Under some approaches, developers may need to utilize multiple stacks when developing software. For example, a developer may develop software on a developer stack, and then deploy the software to a production stack. As development progresses on a stack (e.g., a development stack), ontologies associated with the stack may be modified. Updating ontologies of corresponding stacks (e.g., a production stack) can be time consuming and computationally expensive. These and other drawbacks exist with conventional data management systems.